Designing Personality-Adaptive Conversational Agents for Mental Health Care

Abstract

Millions of people experience mental health issues each year, increasing the necessity for health-related services. One emerging technology with the potential to help address the resulting shortage in health care providers and other barriers to treatment access are conversational agents (CAs). CAs are software-based systems designed to interact with humans through natural language. However, CAs do not live up to their full potential yet because they are unable to capture dynamic human behavior to an adequate extent to provide responses tailored to users’ personalities. To address this problem, we conducted a design science research (DSR) project to design personality-adaptive conversational agents (PACAs). Following an iterative and multi-step approach, we derive and formulate six design principles for PACAs for the domain of mental health care. The results of our evaluation with psychologists and psychiatrists suggest that PACAs can be a promising source of mental health support. With our design principles, we contribute to the body of design knowledge for CAs and provide guidance for practitioners who intend to design PACAs. Instantiating the principles may improve interaction with users who seek support for mental health issues.

Full text

Vol.:(0123456789)
1 3
https://doi.org/10.1007/s10796-022-10254-9
Designing Personality-Adaptive Conversational Agents forMental
Health Care
RanginaAhmad1 · DominikSiemon2· UlrichGnewuch3· SusanneRobra‑Bissantz1
Accepted: 19 January 2022
© The Author(s) 2022
Abstract
Millions of people experience mental health issues each year, increasing the necessity for health-related services. One emerg-
ing technology with the potential to help address the resulting shortage in health care providers and other barriers to treatment
access are conversational agents (CAs). CAs are software-based systems designed to interact with humans through natural
language. However, CAs do not live up to their full potential yet because they are unable to capture dynamic human behavior to
an adequate extent to provide responses tailored to users’ personalities. To address this problem, we conducted a design science
research (DSR) project to design personality-adaptive conversational agents (PACAs). Following an iterative and multi-step
approach, we derive and formulate six design principles for PACAs for the domain of mental health care. The results of our
evaluation with psychologists and psychiatrists suggest that PACAs can be a promising source of mental health support. With
our design principles, we contribute to the body of design knowledge for CAs and provide guidance for practitioners who intend
to design PACAs. Instantiating the principles may improve interaction with users who seek support for mental health issues.
Keywords Personality-adaptive conversational agent· Chatbot· Virtual assistant· Mentalhealth· Artificial intelligence·
Design science research
1 Introduction
The necessity for mental health-related services has
increased rapidly (Luxton, 2020). The prevalence of
mental health issues (such as anxiety, depression or
loneliness) is increasing among people worldwide:
One in four people in the world is affected by mental
health issues at some point in their lives (WHO, 2017).
Access to therapists is limited, resulting in an acute
shortage of mental health workers globally (Ta etal.,
2020; WHO, 2021). Waiting times to receive treatment,
therefore, can profoundly affect a person’s quality of
life (Luxton, 2014). Further barriers to treatment access
include high costs and unequal access to health care
resources, but also attitudinal barriers, such as stigma
toward professional treatments and skepticism about the
effectiveness of treatment (Wasil etal., 2021). Techno-
logical advancements, however, have improved access
to mental health care services (Jones etal., 2014). Con-
versational agents (CAs) represent one emerging tech-
nology with the potential to help address the barriers
that contribute to the unmet health care needs (Luxton,
2020). CAs are software-based systems designed to
interact with humans through natural language (Feine
etal., 2019). CA is the overarching and general term for
software that interacts with users via written or spoken
natural language, and includes systems such as chatbots,
which provide text-based communication, and virtual
* Rangina Ahmad
rangina.ahmad@tu-braunschweig.de
Dominik Siemon
dominik.siemon@lut.fi
Ulrich Gnewuch
ulrich.gnewuch@kit.edu
Susanne Robra-Bissantz
s.robra-bissantz@tu-braunschweig.de
1 Chair ofInformation Management, Institute ofBusiness
Information Systems, Technische Universität Braunschweig,
Mühlenpfordtstraße 23, 38106Braunschweig, Germany
2 Department ofSoftware Engineering, School ofEngineering
Science, LUT University, Mukkulankatu 19, 15210Lahti,
Finland
3 Institute ofInformation Systems andMarketing, Karlsruhe
Institute ofTechnology (KIT), Kaiserstraße 89-93,
76133Karlsruhe, Germany
Information Systems Frontiers (2022) 24:923–943
/ Published online: 2 March 2022
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Information Systems Frontiers (2022) 24:923–943
1 3
assistants, which rely primarily on voice-based interac-
tion (Diederich etal., 2022). Applying Artificial intel-
ligence (AI)-based methods, such as machine learning
and natural language processing (NLP), CAs have rap-
idly become more efficient and enable broad and often
cross-topic conversations (Diederich etal., 2022; Nißen
etal., 2021). A benefit of CAs is that they are accessible
anywhere and available at any time to provide coun-
seling and deliver therapeutic interventions, and they
may therefore be a promising source of support helping
people cope with mental health issues (Luxton, 2020;
Ta etal., 2020). A recent review by Abd-Alrazaq etal.
(2021) demonstrates that the usefulness level of mental
health care CAs is perceived as high by patients and that
they overall have positive perceptions of and opinions
about the CAs. Though AI-based CAs are increasingly
capable of handling highly complex tasks, such as social
support and therapeutic decision-making in a human-
like way, currently available CAs, however, do not live
up to their full potential yet (Fiske etal., 2019; Graham
etal., 2019). Capturing dynamic human behavior to an
adequate extent, and providing responses and reactions
tailored to users’ individual contexts, special interac-
tion dynamics and personalities, still pose a challenge
in designing CAs (Grudin & Jacques, 2019; McTear
etal., 2016). The field of tailored health communica-
tion has long established the need to personalize con-
versation (Abd-Alrazaq etal., 2021; Smith etal., 2015),
since language is a primary tool to understand patients’
experiences and express therapeutic interventions (Lar-
anjo etal., 2018). Psychotherapy, in particular, is highly
patient-centered in clinical practice, requiring skills
such as observing patient behavior and adapting to their
individual personalities and needs accordingly (Graham
etal., 2019; Laranjo etal., 2018). In their interaction
with individuals, contemporary CAs lack the ability to
dynamically change their own personality to adapt to the
user’s personality. Provided a CA has features normally
associated with humans, such as the use of natural lan-
guage or human-like appearance (Nass & Moon, 2000),
human beings tend to treat computer systems as social
entities and ascribe different personality traits to them
(Nass etal., 1994). Thus, as individuals’ interactions
with computers have a fundamental social nature, users
feel more appreciated and comfortable interacting with
the machines when they perceive CAs as more human-
like (Moon & Nass, 1996; Nass etal., 1993). One way
to ensure that a CA is perceived as human-like, is to
provide the CA with the ability to dynamically change
its personality traits. Since personality preferences dif-
fer from person to person (McCrae & Costa Jr, 1997),
the personalization of CA conversation is highly impor-
tant and required (Abd-alrazaq etal., 2019). Based on
an established personality model, and with advances in
NLP, we propose the concept of a personality-adaptive
conversational agent (PACA). PACAs are able to recog-
nize and express personality by automatically inferring
personality traits from users, giving them the ability to
adapt to the changing needs and states of users when
establishing a personalized interaction with them. As
personality differences are manifested in language use,
engagement with users can be further enhanced through
tailored conversation styles (Kocaballi etal., 2020) .
While there is a large body of descriptive knowledge
on design elements or cues that can be adapted, there
is a lack of prescriptive knowledge on ways in which to
actually design PACAs. Therefore, we pose the follow-
ing research question (RQ):
RQ: How to design personality-adaptive conversational
agents (PACAs) to improve interaction with users in men-
tal health care?
To answer our RQ, we conduct a research project fol-
lowing a design science research (DSR) approach (Hevner
etal., 2004). Our design approach is particularly anchored
in two existing kernel theories, namely the five factor
model of personality (McCrae & John, 1992) and the ‘com-
puters are social actors’ paradigm (Nass etal., 1994). On
the basis of these theories, we believe that PACAs have
the potential to improve interaction with users by person-
alizing their health care (Luxton, 2014). To the best of
our knowledge, there is no study that rigorously derives
requirements from both literature issues and user stories
to develop design principles, and that further evaluates the
preliminary design principles by means of expert inter-
views, in order to propose an expository instantiation that
translates the abstract knowledge, captured in the design
principles, into applicable knowledge. Our results indicate
that practitioners (e.g., designers, developers, etc.) who
instantiate our design principles to create a PACA, receive
a promising level of guidance to improve interaction with
users in mental health care.
The remainder of this paper is structured as follows: In
the section on related work, we give a brief overview of
the history and current developments of mental health care
CAs. We also provide an overview of current research and
highlight the critical aspects in the use of mental health
care CAs. In section three, we illustrate the theoretical
foundations of our DSR project. Section four presents our
research methodology in more detail. Section five contains
our derived and evaluated design principles for PACAs,
as well as a demonstration of an expository instantiation.
Finally, we discuss our results, current limitations, and
contributions of our work, and close with a conclusion.
924
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Information Systems Frontiers (2022) 24:923–943
1 3
2 Related Work
Efforts to develop software-based systems within the
health care environment have a long history. In fact,
ELIZA – widely considered the first CA in history – first
appeared in 1966 in a psychotherapeutic context (Weizen-
baum, 1966). Though only developed for demonstration
and not commercial purposes, the simple computer pro-
gram was able to mimic a Rogerian psychotherapist com-
municating with people in an empathic manner (Weizen-
baum, 1966). The interaction resulted in people ascribing
human characteristics to it, and some psychiatrists seeing
ELIZA’s potential computer-based therapy as a “form of
psychological treatment” (Kerr, 2003, p. 305). The under-
lying technology of ELIZA was rather simple: By search-
ing the textual input of its conversation partner for relevant
keywords, the machine produced appropriate responses
according to rules and directions based on scripts hand-
crafted by the programmers (Natale, 2018; Peters, 2013).
PARRY, another early example of a prototype CA devel-
oped in 1979, was designed to simulate and behave like a
person with paranoid schizophrenia (Shum etal., 2018).
The developers’ intention was to find out if other psychia-
trists could determine a real paranoid patient from their
computer model (Shah etal., 2016). PARRY was a rule-
based CA and worked in a similar way as ELIZA, though
with better language understanding capabilities (Shum
etal., 2018).
Current interest in CAs for mental health care is also
nascent, as can be seen in the growing number of online
services offered by health care providers (Bendig etal.,
2019). More than one-fourth of 15,000 mobile health apps
focus on mental health diagnosis or support, according to
the World Health Organization (Abd-Alrazaq etal., 2021).
Over the last few years, three particularly prominent thera-
peutic mental health CAs based on AI technologies have
emerged: Woebot (Woebot Health, 2021), Wysa (Wysa,
2021) and Tess (X2, 2021). These CAs are publicly avail-
able mobile phone applications aimed at helping people to
manage symptoms of anxiety and depression by providing
counseling services. Woebot is a CA that is built to assess,
monitor and respond to users dealing with mental health
issues (Woebot Health, 2021). The CA has a responsive
way to intervene with in-the-moment help and provide
targeted therapies based on cognitive behavioral therapy
(Woebot Health, 2021). The AI chatbot Wysa is also based
on cognitive behavioral therapy, but employs several other
methods, such as behavioral reinforcement and mindful-
ness, to help clients with depression (D’Alfonso, 2020;
Wysa, 2021). According to its developers, Wysa provides
24/7 high-quality mental health support (Wysa, 2021).
The mental health chatbot Tess pursues a similar approach
by being available 24/7 and delivering conversations in
the same way that “a coach or friend would” (X2, 2021).
Preliminary studies of the efficacy of all three applica-
tions have shown a significant reduction in depression
and anxiety levels in the group of participants using the
CAs (D’Alfonso, 2020). Though the presented CAs were
developed to have specific personalities, they are not able
to dynamically change these or to infer user personality in
order to be more adaptive towards the users’ needs. There-
fore, all three CAs represent a rather “one-size-fits-all”
solution, by not adequately adapting to the specificities
of their users.
Ever since research in the field of human-machine inter-
action stressed the importance of avoiding “one-size-fits-all”
interactions, the customization or personalization of CAs to
individual users have become an important research topic
(Abd-Alrazaq etal., 2021). Though this topic is still in its
infancy (Kocaballi etal., 2019), there has been an increased
interest in studies addressing the aspect of personality adap-
tivity in CAs. Studies from Yorita etal. (2019), Kampman
etal. (2019) and Ahmad etal. (2020a, b) show that it is
technically feasible to develop a CA that adapts its own
personality traits to match the identified traits of the user.
Yorita etal. (2019) developed their CA specifically with
the purpose of providing emotional support for the user.
While the studies by Yorita etal. (2019), Kampman etal.
(2019) and Ahmad etal. (2020a, b) primarily focus on text-
based human-machine interaction, the work of Völkel etal.
presents approaches to adapt a voice assistant’s personal-
ity to the user, in order to improve interaction experience
(Völkel etal., 2020, 2021). Further studies by Ranjbartabar
etal. (2018) and Zalake (2020) particularly deal with CAs
that, among other factors, aim to adapt to user personal-
ity to reduce study stress (Ranjbartabar etal., 2018) or to
promote anxiety coping strategies among college students
(Zalake, 2020). The research of Wibhowo and Sanjaya
(2021) concentrate on CAs for use in clinical psychology
and psychotherapy: The authors developed a CA as a warn-
ing system to prevent individuals with borderline personal-
ity disorder from committing suicide. The previous studies
mainly employ empirical methods to describe behaviors in
interaction with these systems and consequently generate
descriptive knowledge about the use and effectiveness of
personality-adaptivity CAs. In addition, these studies all
have different focuses, so there is still a research gap in pro-
viding prescriptive knowledge about how to design a PACA
to improve interactions with users in mental health care.
Due to technological advancements, AI-based CAs have
become increasingly capable of handling highly complex
tasks with human qualities such as a higher autonomy of
decision-making (Brendel etal., 2021) or expressing human-
like feelings (Porra etal., 2019). Consequently, the applica-
tion of CAs can have diverse impacts on individuals – both
925
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Information Systems Frontiers (2022) 24:923–943
1 3
positive and negative. The field of mental health care raises
ethical considerations by its very nature (D’Alfonso, 2020).
Voices have therefore emerged from the research field asking
for a reassessment of the potential “dark sides” of AI and the
ethical responsibilities of developers and designers (Brendel
etal., 2021; Porra etal., 2019). Critics specifically stress the
caveats of creating and perfecting human-like CAs with sim-
ulated feelings, without considering long-term consequences
for human beings, such as deep, emotional attachments
(Ahmad etal., 2021a). However, human beings treat com-
puter systems as social entities and ascribe different person-
ality traits to them (Nass etal., 1994); hence, they feel more
appreciated and comfortable interacting with the machines
when they perceive CAs as more human-like (Moon & Nass,
1996; Nass etal., 1993). To ensure an authentic interac-
tion experience, CAs therefore should be imbued with some
degree of human-like features(Gnewuch etal., 2017). While
advanced CAs are able to simulate conversations employ-
ing therapeutic techniques, it is “not on the near horizon”
(D’Alfonso, 2020, p. 113) for CAs to replicate human thera-
pists. In fact, researchers agree that mental health care CAs
should be used primarily as support systems, since the inter-
action experience and relationship that develops between a
therapist and a patient is considered a significant factor in
the outcome of psychological therapy (D’Alfonso, 2020) and
cannot easily be substituted by a machine. The role of CAs in
mental health care, rather, is to address individuals in need
of treatment who are not receiving any treatment at all due to
various barriers (Bendig etal., 2019; Stieger etal., 2018). In
this way CAs could provide low-threshold access to mental
health care but also bridge the waiting time before approval
of psychotherapy (Bendig etal., 2019; Grünzig etal., 2018).
Mental health care CAs have the potential to create their own
form of interaction experience with users, provided the CA
responses to users are tailored to their individual personali-
ties. As a result, a personalized interaction would improve
health outcomes among care seekers (Luxton, 2014). Given
the widespread interest in mental health care CAs and the
lack of prescriptive knowledge on how to design PACAs,
it is important to address this research gap for researchers,
designers, developers and clinicians.
3 Theoretical Background
3.1 Computers Are Social Actors
Extant research has shown that humans treat computers as
social actors (Nass etal., 1994). The underlying reason,
according to the ‘computers are social actors’ paradigm,
is that humans automatically respond to social cues from
computers (e.g., human-like avatars) in ways similar to
how they would respond to social cues from another person
(e.g., facial expressions, gestures) (Nass etal., 1994). This
behavior can also be observed when users interact with
CAs (Schuetzler etal., 2018). The fundamental reason is
that CAs engage with users in a uniquely human activity,
that is, having a conversation in natural language. Accord-
ing to Fogg (2002), interactive language use is one of the
most salient social cues triggering social responses from
users. In addition, CAs can display many other social cues,
such as human-like avatars, names, vocalization or gestures
(Feine etal., 2019). Research has shown that when users
respond to these social cues from a CA, they perceive it
as more human-like and feel more comfortable interacting
with it (Moon & Nass, 1996; Nass etal., 1993). Studies
have also shown that incorporating social cues can posi-
tively influence various CA-related outcomes, such as trust,
enjoyment, and satisfaction (Kocaballi etal., 2019; Liu &
Picard, 2005; Schuetzler etal., 2018). And since human
beings treat computer systems as social entities and ascribe
different personality traits to them (Nass etal., 1994), they
perceive CAs as more human-like, particularly when the
CAs’ expressions are attuned to the users’ state (Liu & Pic-
ard, 2005). For example, depending on the strength of a CA’s
language, on the expressed confidence level, as well as on
the interaction order, participants attribute an extraverted
or introverted personality to the CA (Moon & Nass, 1996;
Nass etal., 1995). Current CAs are designed to have a pre-
defined personality (e.g., extraverted) and thus are not able
to dynamically change their personality traits based on who
they are interacting with. However, research has shown that
users’ individual personality differences interact with the
CA’s personality (Al-Natour etal., 2005; Gnewuch etal.,
2020). As a result, different users may prefer different CA
personalities, indicating that personality-adaptive CAs may
be a sensible choice to meet user needs and preferences.
3.2 The Five Factor Model ofPersonality
Personality is loosely defined as the construct that differenti-
ates individuals from each other, but at the same time makes
a human being’s behavior, thoughts and feelings (relatively)
consistent (Allport, 1961). The dispositional approach con-
siders trait as the key concept of the field of personality.
In order to measure an individual’s personality, a widely
used classification of personality – the five factor model or
“Big Five” – has been applied in research (McCrae & John,
1992). Compared to other existing personality models, this
multifactorial model was found to be stable across cultures
and observers, and provides a taxonomy for the systematic
evaluation of individuals (Goldberg, 1993; McCrae & John,
1992). The five fundamental traits that have been identi-
fied in this context are openness (to experience), neuroti-
cism (also known as emotional range), conscientiousness,
agreeableness, and extraversion (McCrae & John, 1992).
926
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Information Systems Frontiers (2022) 24:923–943
1 3
The five factor approach to the taxonomy of traits is based
on natural language, and more precisely lexical resources
(Bouchet & Sansonnet, 2012). The lexical hypothesis states
that most socially relevant and salient personality character-
istics have become encoded in the natural language (Boyd &
Pennebaker, 2017). Thus, language (verbal and non-verbal)
has been found to be a fundamental dimension of personality
(Boyd & Pennebaker, 2017). Human language reflects the
psychological state and personality based on the frequency
with which certain categories of words are used, as well as
on the variations in word usage (Boyd & Pennebaker, 2017;
Golbeck etal., 2011; Yarkoni, 2010). Psychologists have
documented the existence of such cues by discovering cor-
relations between a range of linguistic variables and person-
ality traits across a wide range of linguistic levels (Mairesse
etal., 2007). Language use has furthermore been scientifi-
cally proven to be unique, relatively reliable over time, and
internally consistent, and as Boyd and Pennebaker (2017, p.
63) further state: “Language-based measures of personality
can be useful for capturing/modeling lower-level personality
processes that are more closely associated with important
objective behavioral outcomes than traditional personal-
ity measures.” In addition to a speaker’s semantic content,
utterances convey a great deal of information about the
speaker, and the more extreme a person’s personality trait,
the more consistently that trait will be a factor in their behav-
ior (Mairesse & Walker, 2010). For example, extraverts
have been found to have a higher rate of speech, to speak
more, louder, and more repeatedly, with fewer hesitations
and pauses, have higher verbal output, and use less formal
language, whereas people who are highly agreeable show
a lot of empathy, agree and compliment more, use longer
words and many insight words, and make fewer personal
attacks on their interlocutor (Mairesse & Walker, 2010). The
combination of personality-specific words that people use
in everyday life are internally consistent, vary considerably
from person to person and is predictive of a wide range of
behaviors (Boyd & Pennebaker, 2017; Pennebaker, 2011).
Current personality mining services apply AI, specifically
NLP technologies, to automatically infer personality traits
from an individual’s speech or text (Ferrucci, 2012). NLP
techniques particularly subserve to the interpretation of mas-
sive volumes of natural language elements by recognizing
grammatical rules (e.g., syntax, context, usage patterns) of
a word, sentence or document (Ferrucci, 2012).
4 Methodology
To generate design knowledge about the design of PACAs,
we follow the design science research (DSR) approach pro-
posed by Hevner etal. (2004). The overall goal of DSR pro-
jects is to generate rigorously derived and relevant design
knowledge. Design knowledge thereby covers all aspects of
the relationship between the problem and the solution space
(Venable, 2006; Vom Brocke etal., 2020), and aims to create
prescriptive knowledge that contributes to both the theory
and practice of solving real-world problems (Hevner, 2020).
By providing prescriptive solution-oriented design knowl-
edge, DSR therefore represents a complement to descriptive
research that explains the nature of a phenomenon or prob-
lem. Gregor and Hevner (2013) define descriptive knowl-
edge as omega-knowledge, which includes “descriptions
of natural, artificial, and human-related phenomena” and is
“composed of observations, classifications, measurements,
and the cataloging of these descriptions into accessible
forms” (p. A2). In contrast, DSR also creates prescriptive
knowledge about artifacts that address the phenomenon or
problem (Gregor etal., 2007), which Gregor and Hevner
(2013) define as lambda-knowledge. DSR should not only
generate original solutions to existing problems, but also
demonstrate the beneficial application of such solutions to
the problem and show their broad implications (Baskerville
& Pries-Heje, 2010; Vom Brocke etal., 2020), thus gener-
ating both omega and lambda-knowledge (Hevner, 2020).
Prescriptive design knowledge (lambda-knowledge) can
be represented in different ways, such as design principles
(DPs), design features, or instantiations (Hevner, 2020;
Vom Brocke etal., 2020), whereas descriptive knowledge
(omega-knowledge) describes observations, measurements
or classifications of phenomena and sense-making, such as
patterns or natural laws (Gregor & Hevner, 2013). Hence,
DSR projects involve an interplay of synergies between
descriptive and prescriptive knowledge, as well as the use
of existing knowledge and the production of new knowledge
(Hevner, 2020).
DPs as a form of formalized knowledge are gaining popu-
larity in the field of information systems (IS) because they
allow researchers to capture abstract or meta-knowledge that
addresses a class of problems, rather than a single problem
(Gregor etal., 2020; Iivari etal., 2021; Purao etal., 2020).
To ensure practical relevance and applicability, DPs should
imply accessibility, effectiveness, and, most importantly,
guidance for action (Iivari etal., 2021). In our research, we
follow an iterative and multi-step process that is fundamen-
tally embedded in the three-cycle view of DSR (i.e., rele-
vance cycle, rigor cycle and design cycle) (Hevner, 2007), to
generate design knowledge for the design of PACAs for men-
tal health care. The relevance cycle incorporates issues from
the problem space in the study and places outcomes from
the design cycle in evaluation and practice. The rigor cycle
involves the use of kernel theories and scientific methods, as
well as research experience and expertise, in the study, while
also adding new knowledge created by the researchers to the
growing knowledge base. The design cycle, as the core of
DSR, includes construction activities for the development
927
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Information Systems Frontiers (2022) 24:923–943
1 3
and evaluation of artifacts and design processes (Hevner,
2007). In our DSR project, we generate design knowledge in
the form of DPs and an expository instantiation, by involving
all three cycles multiple times. We follow the first strategy
as proposed by Iivari (2015), in which we construct a meta-
artefact (i.e., DPs) as a general solution concept, and then
further adapt and follow the methodological components
proposed by Möller etal. (2020) for the development of DPs
in IS. These authors propose a supportive approach as well
as a reflective approach that differ within “the point of arti-
fact design and the logic of generating design principles” (p.
214). The two approaches of Möller etal. (2020) are similar
to the two general strategies proposed by Iivari (2015) in
that either generalizable design knowledge is created first,
which is then concretized, or concrete design knowledge
(e.g., design features or instantiations) is created first, which
is then abstracted and generalized. However, the approach of
Möller etal. (2020) focuses specifically on the construction
of DPs and proposes a methodological procedure. We follow
the supportive approach, in which DPs are “the provision
of design knowledge in advance to support the design of
an artifact before the design process takes place” (p. 214).
Figure1 shows our DSR approach based on Hevner (2007),
Iivari (2015) and Möller etal. (2020).
Our DSR approach is divided into two main phases,
addressing the problem space and the solution space (includ-
ing the evaluation of the solution), which in turn consist
of three sub-steps, respectively. To generate and evaluate a
solution in the form of DPs for the identified problems, we
use a mapping diagram. Mapping diagrams help to visualize
the connection and derivation logic between DPs and meta-
requirements, as well as between the problem space and the
solution space (Möller etal., 2020). In this way, connections
between and derivations of the individual aspects become
clearer. The mapping diagram of the derivation and con-
struction of the MRs and DPs can be found in the Appendix
(see Fig.4). An interplay of synergies between the rigor
cycle, relevance cycle, and design cycle is reflected in each
sub-step.
Step 1: Problem Space
An important prerequisite for effective and practical
design knowledge is a good understanding and description of
the underlying problem space. To conceptualize the problem
space, Maedche etal. (2019) propose considering and ana-
lyzing four key concepts of the problem space; stakeholders,
needs, goals, and requirements. In our approach, the process
of understanding and defining the problem space consists
of three steps addressing the needs of the stakeholders and
issues from the application domain, which we subsequently
captured in requirements. To conceptualize the problem
space, we followed three steps that build upon each other:
a) Identifying issues from the application domain (rel-
evance cycle), literature and kernel theories (rigor cycle)
b) Deriving user stories (relevance cycle) through a qualita-
tive survey (rigor cycle)
c) Deriving meta-requirements (design cycle) through user
stories, literature issues, application domain aspects, and
kernel theories
We approached the problem space by reviewing extant
literature and identifying current problems in mental health
Fig. 1 Research approach based
on Iivari (2015), Möller etal.
(2020) and Hevner (2007)
928
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Information Systems Frontiers (2022) 24:923–943
1 3
care. In a next step, and based on our kernel theories, we
conducted a user survey to capture the perspective of poten-
tial users and to derive meta-requirements together with the
literature issues. We followed an explorative approach and
conducted a qualitative study (Babbie, 2020) using an open
questionnaire with the aim to capture comprehensive user
stories about PACAs in the context of mental health care.
Our open questionnaire consisted first of an extensive expla-
nation of the functionality and nature of a PACA to make
sure participants from all backgrounds understood the con-
cept. After that, we used an open questionnaire to capture
user stories about the design of PACAs in the context of
mental health care. These open questions asked for general
requirements (e.g., mental health support, safety and pri-
vacy) and design-related requirements (e.g., behavior, com-
munication style) (see Table1 for an excerpt from the survey
with sample responses). To help participants visualize what
a conversation between a patient and a PACA might look
like, we created a predefined chat record. For our simulated
dialogue, we used the conversational design tool Botsociety,
which allows prototyping and visualizing CAs. The conver-
sation was provided in the form of a video. Figure2 shows
a mock-up of the simulated interaction. The video with the
entire conversation can be viewed here: https:// youtu. be/-
sfSNJ wCCI0
The survey was distributed via our private network and
the crowdsourcing platform Mechanical Turk (mTurk) and
was carried out in December 2020. A total of 60 respondents
participated in the study, producing more than 6865 words
of qualitative data answering the open questions, which took
roughly between 25 and 40min to complete. Table1 shows
an excerpt from the survey with four open questions and
example responses that were used for qualitative content
analysis.
Participants (32 male, 28 female) were between 23 and
71years old, with an average age of 36years. In order to
analyze the data, we followed a qualitative content analysis
by coding the answers of the participants, which consisted
mainly of inductive category forming (Mayring, 2014). The
authors conducted the coding process independently, and
whenever the results differed, we discussed the points of
disagreement until we reached a consensus. In the last sub-
step, we constructed meta-requirements from the captured
user stories, the issues from the literature and the application
domain.
Step 2: Solution Space and Evaluation
Our proposed solution to the problem space identified in
step 1 is a PACA for mental health care, for which we pre-
sent design knowledge in the form of DPs. We constructed
our DPs based on the meta-requirements, and then evaluated
them with experts from the application domain. For the sys-
tematic derivation and evaluation of our solution space, we
followed the following steps:
a) Constructing design principles (design cycle).
b) Evaluating design principles (design cycle) with expert
(relevance cycle) interviews (rigor cycle).
c) Designing an expository instantiation (design cycle).
Our DPs are formulated based on the approach proposed
by Gregor etal. (2020, p. 2), who defined the anatomy of
design principles so that DPs are “understandable and use-
ful in real-world design contexts.” The authors point out
the importance of including the concerned actors when
formulating the DPs to complement the anatomy, with the
aim that DPs should be “prescriptive statements that show
how to do something to achieve a goal” (Gregor etal.,
2020, p. 2). The anatomy of a DP consists of the aim,
implementer, and user; context; mechanism; and ration-
ale, and is presented in the form of heuristic rules. The
Table 1 Excerpt from the survey with sample responses
Question Example response
Do you think the concept of a PACA - with the computer system hav-
ing a personality and being able to adapt to the user’s personality - is
useful/ helpful in mental health therapy?
“Yeah I think it makes you feel less like you’re just sharing your emo-
tions with an inanimate object, it makes you feel like there is some
more meaning to sharing it with something that can at least pretend to
care. It makes it feel more worthwhile to have it respond with some-
thing other than just a planned response.”
Please comment briefly, why you think a specific trait is important in
your opinion.
“Being talkative will help in communicating feelings and affection
is a trait that people with mental problems struggle to receive and
comprehend.”
In your opinion, in which of the roles should a mental healthcare
PACA slip into? Please explain briefly why.
“Looking at mental states means an emotional state of mind. A friend or
companion are the closest to a patient and are therefore necessary.”
What are, in your opinion, reasons that speak against communicating
with a PACA? What concerns would you have in your interaction
with the PACA?
“I suppose it raises questions or suspicions about who could be reading
or observing your conversations with the PACA. How secure and con-
fidential is your information in the short and long term. It’s a question
of trust.”
929
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Information Systems Frontiers (2022) 24:923–943
1 3
evaluation of constructed artifacts such as DPs is an essen-
tial step in the design cycle to generate rigorous design
knowledge. To evaluate our constructed DPs, we chose a
qualitative approach in the form of expert interviews. To
ensure relevance to the application domain, we selected
experts, such as psychologists and psychiatrists, and had
them evaluate the constructed DPs from a psychothera-
pist’s perspective. The criteria for selecting experts were
a minimum of at least two years of professional experience
in the care or therapy of mentally ill people, as well as a
completed apprenticeship or studies in medicine, psychol-
ogy, or another related field. The experts were selected and
approached from our personal network. The work of EX1
(a psychologist) and EX6 (a social worker and therapist)
particularly focus on youth care, while EX2 to EX5 are
all psychiatrists who trained as psychotherapists and work
with patients of all ages. EX2 deals specifically with geri-
atric psychiatry. The six experts were interviewed between
March and April 2021. Each interview took between 50
and 80min. Table2 shows an overview of the interview
panel, including the experts’ education and professional
background.
The interviews were conducted with the support of a
semi-structured interview guideline (Mayring, 2014). Inter-
views that follow a guideline are based on pre-defined ques-
tions that help orient the interviewer and thus ensure that all
critical aspects of the scope of the interview are included. In
addition, queries and deviations by the experts are possible
(Mayring, 2014). The interview guide started with general
questions about the profession of the experts, their profes-
sional experience, and their experience in dealing with CAs.
This was followed by a detailed explanation of the concept
of a PACA and follow-up questions by the interviewer to
ensure that the concept was understood. Thereupon, all DPs
were examined individually and evaluated by the experts,
who were asked to recapitulate each DP in their own words,
to again check whether they understood each DP. The
experts were then asked to comment on each DP in terms
of its relevance and to add possible missing aspects or high-
light particularly relevant aspects. Subsequently, the audio
recordings were transcribed and coded using MaxQDA (ver-
sion 2020) qualitative data analysis software. The results
were incorporated in the further development of the DPs.
Despite the anatomy of a DP and different frameworks for
Fig. 2 Mock-Up of the conver-
sation between Raffi (PACA)
and Jules (User)
930
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Information Systems Frontiers (2022) 24:923–943
1 3
the accumulation of design knowledge (Rothe etal., 2020),
the reusability and use of DPs for practitioners is often
challenging. Iivariet al. (2021) address this problem and
specifically propose strengthening the actability and guid-
ance for the implementation of DPs to maintain the practical
relevance of DSR. Since DPs represent formalized design
knowledge, we design an expository instantiation (Gregor
etal., 2007) that translates the abstract knowledge captured
in DPs into applicable knowledge. The next section presents
the result of the two steps and their sub-steps.
5 Designing PACAs forMental Health Care
5.1 Deriving Design Principles forPACAs
Within our problem space, we first identified several issues
specific to our application domain (AD), as well as current
literature issues (LI) related to CAs. The issues relating to
the AD consequently establish the relevance of our research
due to societal needs (i.e., unmet mental health care). The
LIs reveal current issues and related theories of CAs regard-
ing, for example, their interaction capabilities, which are
captured in the literature. In contrast to the issues of the
AD, the LIs originate from the knowledge base and show
us, for example, why CAs have limited conversational abil-
ity. Both thus contribute to the relevance and rigor of our
DSR project.
The World Health Organization reports that one in four
people in the world is affected by mental health issues at
some point in their lives (WHO, 2017). Particularly in times
of humanitarian crisis, the necessity for health-related ser-
vices increases rapidly (Luxton, 2020;Torous etal., 2020
; WHO, 2021) (AD1). As a result of this increase, there
is a serious shortage of mental health workers globally (9
per 100,000 population), which in turn contributes to unmet
health care needs (Luxton, 2020; Prakash & Das, 2020)
(AD2). Thus, there is a high need for offering IT-based men-
tal health services to surpass the availability of healthcare
workers and ease the burden on them. In addition to the high
prevalence of mental health issues, there is a strong social
stigma attached to mental illness (WHO, 2017). Therefore,
patients with mental health issues are considered particularly
vulnerable. Studies also show that the personality of a user
plays a crucial role in the adoption of emerging technology
that raises concerns about data security and privacy (Junglas
etal., 2008). Since the field of mental health handles highly
sensitive data, paying attention to the user’s personality is
important. Hence, if patient safety is not addressed appropri-
ately, a lack of privacy mechanisms, and thus a loss of trust,
and could cause harm to people who exhibit sensitive men-
tal health conditions (Luxton, 2020) (AD3). Although CAs
are considered an emerging technology with the potential
to help address these issues, the adoption of CAs in mental
health care has been slower than expected (Graham etal.,
2020). Psychotherapy is a highly patient-centered clinical
practice. This means that a successful conversation is par-
ticularly dependent on each patient’s individual dynamic
behavior and the therapist’s ability to adapt to the patient’s
specific personality in order to form a therapeutic relation-
ship (Laranjo etal., 2018; Luxton, 2014). However, contem-
porary CAs do not capture user personality and individual
dynamic human behavior to an adequate extent (Ahmad
etal., 2021b; Yorita etal., 2019) (LI1). As a result, the
CA’s knowledge of the patient’s personality and behavior is
restricted, and their ability to effectively adapt and conse-
quently to develop rapport with the patient is limited. The
needs and preferences of users while interacting can be fun-
damentally different. However, many CAs are focused on a
“one size fits all”-approach instead of pursuing personalized
communication. Contemporary CAs therefore insufficiently
tailor responses to patients’ individual contexts, special
interaction dynamics, and personality (Grudin & Jacques,
2019; McTear etal., 2016) (LI2). This in turn results in
Table 2 Expert panel for the DP evaluation
ID Gender Age Education Professional Background
EX1 Female 31 Studied psychology and has five years of working experi-
ence.
Works in the in-patient youth welfare in a therapeutic resi-
dential school with patients with mental disorders.
EX2 Male 38 Studied human medicine, specialized in neurology and
psychiatry. Has nine years of working experience.
Works in a private clinic for psychiatry and psychotherapy
(geriatric psychiatry) as a senior physician.
EX3 Female 29 Studied human medicine, completed an apprenticeship in
psychotherapy. Has four years of experience.
Works as a medical specialist in psychiatry and psychother-
apy in an outpatient and inpatient facility.
EX4 Male 28 Studied human medicine, completed an apprenticeship in
psychiatry. Has two years of experience.
Works in a private clinic for psychiatry and psychotherapy.
EX5 Female 32 Studied human medicine, specialized in psychiatry. Has six
years of experience.
Works in a private clinic for psychiatry and psychotherapy as
a senior physician.
EX6 Female 35 Studied social work, has completed an apprenticeship as a
therapist. Has eight years of working experience.
Works in residential youth care in a therapeutic residential
home with patients with mental disorders.
931
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Information Systems Frontiers (2022) 24:923–943
1 3
decreased communication satisfaction, user engagement, and
dialogue quality (Kocaballi etal., 2019). Another literature
issue concerns current CAs’ limited ability to hold longer
conversations or answer more complex questions (Chakra-
barti & Luger, 2015; Clark etal., 2019). Although many CAs
are increasingly capable of handling highly complex tasks
with their natural language understanding capabilities, CAs
are not sophisticated enough to recreate the richness of a
conversation with a human therapist (Gaffney etal., 2019)
(LI3). As a consequence, CAs fail to sufficiently engage
the patients and to respond with suitable advice (Kocaballi
etal., 2019). A further identified literature issue is a CA’s
degree of anthropomorphism: The perceived level of human-
ness can either have a positive or a negative impact on users
(Prakash & Das, 2020) (LI4). Hence, the degree of anthro-
pomorphism of a CA can affect a patient’s judgments and
attitudes (Kim etal., 2019), which can lead to the uncanny
valley phenomenon - a feeling of uncanniness when CAs
become too human-like (Kim etal., 2019). Table3 sums up
all ADs and LIs.
Our underlying kernel theories – the five factor model and
the ‘computers are social actors’ paradigm – are reflected in
the user stories (US) derived from our qualitative study (see
1.b) of Fig.1 and Table1 for example responses). When
the 60 participants recruited using mTurk and from our pri-
vate network were asked to describe their ideal PACA for
mental health care, they ascribed the virtual therapist with
similar social cues and characteristics as they would have
a human therapist. We aggregated the most common and
repetitive USs into three categories: support, safety, and
behavior. The category support sums up the ways in which
the PACA could actively support the patient by improving
its mental health service. The USs include around the clock
availability (US1), an easy conversation while interacting
(US2, US3), memorized conversations (US4), a personalized
interaction experience (US5), competence (US6), support
and helpfulness for mental health therapy (US7, US8) and
the ability to refer patients to human therapists (US9). The
category safety contains USs in which many participants
stated that they want the PACA to be trustworthy (US10),
so they can build a relationship with the PACA (US11). A
basic requirement for this would be that the patient’s data
should be secure and not misused by the PACA (US12). The
participants further indicated that there should be an option
where the PACA-patient conversation is monitored by a
human therapist (US13) so that the patient does not become
dependent on the PACA and/or gets desocialized (US14). A
PACA should also be able to notice warning signals (US15).
The third category, behavior, refers to the characteristics
that an ideal mental health care PACA should have, accord-
ing to our participants. Although many of the participants
agreed that a PACA should be and act rather human-like
(US16), as well as be able to communicate via voice and
facial expressions (US17), their preferences concerning the
PACA’s communication style and personality traits varied
strongly between participants. US18-US28 in Table Table4
sums up the aggravated PACA characteristics that were men-
tioned at least twice by various participants.
Based on our understanding of the problem space (i.e.,
AD, LI, US), we derived seven meta-requirements (MR)
that ultimately led to six DPs for PACAs. After our expert
evaluation, we adjusted the wording and refined the prin-
ciples following the structure of Gregor etal. (2020). We
elaborate on the MRs, as well as on the derived DPs, in
the following section.
Mental health issues are increasing among people
worldwide (AD1) and there is a dramatic shortage of men-
tal health professionals (AD2), which is why users asked
for an easily accessible and always available option (US1)
to support them with their mental health issues (US3). To
address this lack of human support for people in need,
PACAs should always be available for their users, to put
them at ease at any time (MR1). However, even when the
PACA is available 24/7, not all users may actively seek
help when they need it the most (AD3), because mental ill-
ness is still often deemphasized or stigmatized. Users also
stated that they want a supportive and motivating PACA
(US7) that supports them proactively (US24). To prevent
users suffering in silence, PACAs should therefore take
the initiative to reach out to their users on a regular basis
(MR2). Based on these two MRs, we thus propose:
Table 3 Derived issues from application domain & literature
# Application Domain (AD) and Literature Issues (LI)
AD1 Humanitarian crisis: Increased necessity for health-related services (Torous etal., 2020; WHO, 2021)
AD2 Shortage in health care providers and therapists (Luxton, 2020; 2020; WHO, 2021)
AD3 Patient safety: sensitive data and privacy concerns (Luxton, 2020, WHO, 2017)
LI1 Inability to adapt dynamically to a user’s personality (Ahmad etal., 2021b; Yorita etal., 2019)
LI2 No tailored responses to a user’s preferred communication style (Kocaballi etal., 2019)
LI3 Limited conversational ability (Chakrabarti & Luger, 2015; Clark etal., 2019)
LI4 Lacking adaptability to a user’s preferred degree of CA humanness (Kim etal., 2019)
932
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Information Systems Frontiers (2022) 24:923–943
1 3
DP1: Principle of Proactive Support
For designers and developers to design personality-
adaptive conversational agents (PACA) that provide
mental health services to socially support patients
independently of therapist, location and time, ensure
that the PACA is accessible 24/7 and proactively
checks in on the user on a regular basis, so they can
receive support at any time.
Another MR we derived from the problem space is a
PACA’s communication competence, particularly when it
comes to developing and building a long-term relationship
with a patient to improve well-being. Currently, CAs have
limited conversational abilities (LI3), which is why users
expressed their wishes that PACAs should be competent
communication partners (US6) that can memorize past con-
versations (US4), have a rich vocabulary (US19), and are
overall helpful for a patient’s mental health therapy (US8)
by finding supportive and motivating words for the patient
(US7) (MR3). We therefore propose:
DP2: Principle of Competence
For designers and developers to design personality-
adaptive conversational agents (PACA) that provide
mental health services to socially support users inde-
pendently of therapist, location and time, provide the
PACA with a domain-specific knowledge base and
incorporate therapeutic techniques, so the user feels
understood and perceives the PACA as competent.
The fourth MR specifically addresses patient safety
(AD3). PACAs should protect patient data and always
respect the users’ privacy, in ways similar to how human
therapists build therapeutic relationships (US11) by being
a trusted source of mental health support (US10) and
by remaining strictly confidential (US12). Other safety
requirements for a PACA are that it should support its
users but not make them dependent on it (US14). Besides,
a PACA should recognize critical trigger words (e.g.,
suicidal thoughts) and involve a human therapist (US9,
US13). PACAs also need to analyze their patients’ data
to learn more about their individual needs, preferences
and personalities, to give tailored responses (LI2). In this
Table 4 Derived user stories
from qualitative study Category # User Story (US)
Support US1 I want the PACA to be accessible and available for me 24/7.
US2 I want the PACA to be easy to talk to like a friend/pen-pal/therapist.
US3 I want the PACA to be able put me at ease.
US4 I want the PACA to memorize our conversations.
US5 I want a personalized interaction experience with the PACA.
US6 I want the PACA to be competent.
US7 I want the PACA to be supportive and motivate me.
US8 I want the PACA to be helpful for my mental health therapy.
US9 I want the PACA to be able to refer me to a human therapist.
Safety US10 I want the PACA to be trustworthy.
US11 I want to be able to build a relationship with the PACA.
US12 I want my data to be secure and not misused.
US13 I want the PACA to be able to be monitored by my therapist.
US14 I don’t want to get desocialized and/or become dependent on the PACA.
US15 I want the PACA to notice warning signals.
Behavior US16 I want the PACA to be and act human-like.
US17 I want the PACA to be able to have a voice and have facial expressions.
US18 I (don’t) want the PACA to use emojis.
US19 I want the PACA to have a rich vocabulary.
US20 I (don’t) want the PACA to be formal.
US21 I want the PACA to be calm, patient and receptive.
US22 I want the PACA to be kind and polite.
US23 I want the PACA to be humorous, witty and curious.
US24 I want the PACA to be proactive and interactive.
US25 I want the PACA to be confident and firm.
US26 I want the PACA to be empathetic, sensitive and caring
US27 I want the PACA to be open minded and attentive.
US28 I want the PACA to be flexible and adaptive.
933
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Information Systems Frontiers (2022) 24:923–943
1 3
regard, the PACA should be highly transparent towards its
users (MR4). Hence, we propose:
DP3: Principle of Transparency
For designers and developers to design personality-
adaptive conversational agents (PACA) that provide
mental health services to socially support patients
independently of therapist, location and time, ensure
that the PACA is transparent in communicating patient
safety and privacy issues, so that users trust the PACA
with their health-related concerns and feel safe when
sharing sensitive data.
One important aspect that users mentioned frequently
and couched differently, is the PACA’s social role (US2).
Different people in a patient’s life can provide social sup-
port, including therapists, friends, and family members. The
responses in our study showed that different people pre-
ferred different types of social roles for the PACA (US2):
While some participants indicated that their ideal PACA
should, for example, take on the role of a witty and humor-
ous friend (US23) who interacts in a more informal manner
(US20), others described their PACA as a virtual therapist
who should be more confident and firm (US25), and more
formal in appearance (US20). Therefore, PACAs should be
able to take on a social role (e.g., friend, therapist) based
on the user’s individual needs and preferences (MR5). We
consequently propose:
DP4: Principle of Social Role
For designers and developers to design personality-
adaptive conversational agents (PACA) that provide
mental health services to socially support users inde-
pendently of therapist, location and time, allow users
to choose between different social roles so the PACA
can take on the user’s preferred social role and adapt
to their needs.
Today, CAs can largely be customized only by user
input. Customization is also mostly limited to simple
and external aspects of a CA (LI4). However, the per-
ceived level of humanness, which can be expressed
through verbal and non-verbal language, can have
either a positive or negative influence on users (LI4).
A large number of our participants, though, stated that
they wish to have a personalized interaction with their
PACA (US5) and want their PACA to be and act human-
like (US16). They further expressed their wish to inter-
act with a PACA that can communicate non-verbally
via voice and/or facial expressions (US17). Therefore,
adapting the degree of adaptable anthropomorphism of
a PACA, as well as further options of communication
via voice and/or facial/body expressions, is required
(MR6). We therefore propose:
DP5: Principle of Anthropomorphism
For designers and developers to design personality-
adaptive conversational agents (PACA) that provide
mental health services to socially support users inde-
pendently of therapist, location and time, allow the
users to choose what type of PACA they want to inter-
act with (chatbot, voice assistant, embodied conversa-
tional agent), so they can determine the PACA's degree
of anthropomorphism based on individual needs.
Successful psychotherapy depends on the therapist’s
ability to adapt to the patient’s specific personality to form
a therapeutic relationship (LI1–2). However, current CAs
follow a “one size fits all” approach instead of pursuing
personalized communication. This is in line with our study
participants’ requirements, that is, a personalized interaction
experience with the PACA (US5). In addition, users may
have vastly different preferences for communication with
a PACA. This observation aligns with the responses from
our user stories (US18-US28). For example, while some
participants preferred texting with an extraverted and witty
PACA (US23), others indicated they would rather trust an
introverted PACA that communicates in a calm and soothing
voice (US21). In summary, our participants stated that they
not only wish a personalized interaction (US5) but also a
flexible and adaptive PACA (US28) Therefore, the personal-
ity of a PACA should be aligned with the users’ preferences
(MR7). We propose the following principle:
DP6: Principle of Personality Adaptivity
For designers and developers to design personality-
adaptive conversational agents (PACA) that provide
mental health services to socially support users inde-
pendently of therapist, location and time, imbue the
PACA with language cues specific to different per-
sonality dimensions, to enable the PACA to adapt to
the user’s preferred communication style and increase
interaction quality.
5.2 Evaluating Design Principles forPACAs
DP1: Principle of Proactive Support was rated as an impor-
tant principle by all experts; both constant availability and
regular, proactive check-ups were named as highly useful
features for a PACA. EX5 and EX6 pointed out that around-
the-clock availability and more regular check-ins are advan-
tages that PACAs might have over human mental health pro-
fessionals. As to whether the PACA should always act in
a proactive manner, the experts almost unanimously stated
that the majority of patients would most probably appreciate
such a function. Specifically, patients with certain mental
health conditions, such as anxiety disorders or depressions,
are often not able to actively seek help, which is why a
regular check-in by the PACA would be helpful. However,
934
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Information Systems Frontiers (2022) 24:923–943
1 3
EX4 also noted that for serious psychiatric illnesses, such as
schizophrenia or paranoia, a proactive PACA might be too
intrusive, giving the patient the feeling of being observed
and monitored in a way which would be not beneficial for
their mental health state. EX3 and EX5 further indicated that
the patient, or even their treating therapist, should be able
to select the intervals at which the PACA should become
proactive, as too much or too little intervention can influence
the quality of the support. We therefore slightly adjusted the
description for DP1 and now mention that the PACA should
check in on the user on a mutually agreed regular basis (see
Table5).
The experts considered DP2: Principle of Competence
as equally important as DP1, stating that a PACA that is
used in a mental health care context most definitely needs to
have domain-specific knowledge and should be able to use
therapeutic techniques in conversation with the patient. EX3,
EX4 and EX5 remarked that the inclusion of ‘therapeutic
techniques’ does not necessarily mean incorporating tech-
niques that are specific to certain schools of psychotherapy
(e.g., cognitive behavioral therapy, psychodynamic psycho-
therapy) in the PACA, but rather that the PACA should able
to understand the patient’s individual concerns and then
dynamically respond with suitable advice. However, four
of the experts were slightly skeptical whether PACAs could
be designed so that they are sophisticated enough to recre-
ate the richness of a conversation with a human therapist.
While EX2 and EX4 highlighted the importance of “reading
between the lines” in order to react to warning signals, EX3
indicated that the PACA’s way of communication should
also be perceived as “authentic” by showing understand-
ing, otherwise patients could easily get disappointed by the
PACA’s lack of perceived competence. The experts agreed
with the formulation of DP2, and therefore we did not
change the description.
All six experts considered DP3: Principle of Transpar-
ency a fundamental prerequisite for ensuring patient safety.
According to EX1 and EX6, a PACA should be under a duty
of professional secrecy equivalent to that of a healthcare
professional. EX2 further stated that, in order to build trust
with the patient, a PACA should be absolutely transparent
about what happens with the patient’s data, because, as EX4
further remarked, a PACA would otherwise “risk losing their
patients’ trust.” EX5 pointed out that, although it is crucial
to inform the patient about the medical confidentiality at
the beginning of the very first session, it does not necessar-
ily have to be an integral part of every session. However, a
PACA should always be able to address the privacy terms
when asked by the user. EX4 suggested reminding the user
at regular intervals that writing with the PACA is a “safe
space.” Moreover, all experts highlighted the importance
of building a steady therapeutic relationship, as rapport and
trust can only be built over a longer period. To achieve this,
a PACA must be considered a safe space for the users. The
wording of DP3 was accepted by the experts, hence nothing
was changed in the description.
The preliminary version of DP4: Principle of Social Role
was not formulated clearly enough, since the majority of
experts needed some explanation. It was not intuitively clear
what we meant by social role, so we added specific examples
(“e.g., friend, therapist etc.”) to the DP4 description. Once
we mentioned the examples, the meaning was clear and the
experts did not need further explanations. The principle was
rated by the experts as valuable, however not as important
as the previous DPs. EX3 and EX1 mentioned the impor-
tance of specific roles that people reflect for patients and
approved the idea of having a PACA that reflects a specific
role. EX4 and EX6 argued that, from a therapeutic perspec-
tive, it might not always be effective for the patients’ therapy
progress if the PACA continuously takes on the role of a
friend who never “counters or addresses unpleasant topics.”
EX2 further indicated that specific roles or genders can be
associated with fear or aggression. To avoid this, EX5 sug-
gested that the PACA should be able to change its social role
situationally, even within one session. EX2 further assumed
that a patient who can choose between different social roles
for their PACA would be more likely to use the service than
if the option did not exist. After receiving this feedback, we
elaborated on DP4 by adding that the user can also switch
between certain social roles to promote therapy progress
(see Table5).
The experts rated DP5: Principle of Anthropomorphism
as of similar importance as DP4, stating that it promotes
better adaptability to the individual user. From a psycho-
therapeutic perspective, the principle can be specifically
beneficial when patients “cannot talk about their issues, but
rather prefer to write” (EX2), or when patients “have diffi-
culty reading emotions from text, non-verbal language can
help” (EX4). EX6 further stated that some patients need to
feel a PACA’s social presence, for example in the form of an
embodied CA, to open up and feel comfortable. EX3, how-
ever, doubted the efficacy of this, as she was rather critical
towards CAs that are too anthropomorphized, stating that
it could lead to negative dependencies. She therefore indi-
cated that DP5 must be viewed and designed with caution.
EX1 remarked that the PACA’s level of humanness might
affect how patients perceive the PACA’s competence. EX5
suggested that familiar voices can be helpful in crisis mode
and thus can be considered a useful feature for the PACA.
As DP5 was comprehensible and the experts agreed on the
wording of the description, nothing was modified.
Concerning DP6: Principle of Personality Adaptivity, all
experts pointed out that capturing a patient’s dynamic behav-
ior and individual personality during therapy is an essential
step towards forming a trustful therapeutic relationship.
EX2 and EX3 explained that building rapport with patients
935
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Information Systems Frontiers (2022) 24:923–943
1 3
usually takes several hours of therapy until the patients are
able to slowly open up. In this context, EX4, EX2 and EX1
highlighted the importance of language, and stressed that
the style in which they communicate with their patients,
specifically, plays a crucial role in the quality of their inter-
action. All six experts stated that, in the context of mental
health care, the most valuable feature with which to imbue
a PACA is the ability to capture patient personality, just as a
human therapist would do. EX3 stated that, although thera-
pists have their own unique personality and therapeutic tech-
niques, they very much act (to some degree) like a PACA
by adapting to their patients’ personalities. Therefore, DP6
was rated as particularly important. However, EX4 and EX5
remarked that, for therapeutic purposes, the PACA should
also be able to change its personality and communication
style to a “provocative style” (EX4) to “break the patient
through their reserve from time to time” and to “not become
a people pleaser” (EX5). EX4 added that a PACA should be
able to dynamically change its personality, if the goal is to
achieve therapeutic progress. To make DP6 more compre-
hensible, we changed “preferred communication style” to
“personality” (see Table5). Table5 summarizes the revised
and final six DPs.
In general, the experts agreed that the six DPs cover all
important criteria for designing a PACA for mental health
care. They saw potential in designing PACAs to address the
issues from the ADs; however, they also strongly empha-
sized that experts in psychotherapy should be involved in the
design of mental health care PACAs. They further indicated
that a user’s experienced level of mental health issue(s) can
play an important role in whether a PACA can be effective
or not. The psychiatrists stated in particular that realistically,
a PACA alone would not be able to help users with severe
mental health issues, but could rather be a useful tool for
both the patient and their treating therapist. Though they
were familiar with the terms chatbot and digital assistant,
the experts admitted that visualizing a PACA with therapist-
like capabilities would only be possible if they could “expe-
rience [it] first-hand, to see how it really works” (EX2).
5.3 Expository Instantiation
To show the applicability of our DPs and provide guid-
ance for the implementation of a PACA, we developed an
expository instantiation (Gregor etal., 2007; Iivari etal.,
2021). Design knowledge, especially DPs, tends to be
highly abstract and consequently cannot always be imple-
mented easily and directly. Therefore, we transformed our
defined DPs into an expository instantiation that can assist
in “representing the design knowledge both as an expository
Table 5 Evaluated design principles for PACAs
# Title Design Principles (DPs)
DP1 Principle of Proactive Support For designers and developers to design personality-adaptive conversational agents (PACA) that
provide mental health services to socially support patients independently of therapist, location and
time, ensure that the PACA is accessible 24/7 and proactively checks in on the user on a mutually
agreed regular basis, so they can receive support at any time.
DP2 Principle of Competence For designers and developers to design personality-adaptive conversational agents (PACA) that pro-
vide mental health services to socially support users independently of therapist, location and time,
provide the PACA with a domain-specific knowledge base and incorporate therapeutic techniques,
so the user feels understood and perceives the PACA as competent.
DP3 Principle of Transparency For designers and developers to design personality-adaptive conversational agents (PACA) that
provide mental health services to socially support patients independently of therapist, location and
time, ensure that the PACA is transparent in communicating patient safety and privacy issues, so
that users trust the PACA with their health-related concerns and feel safe when sharing sensitive
data.
DP4 Principle of Social Role For designers and developers to design personality-adaptive conversational agents (PACA) that
provide mental health services to socially support users independently of therapist, location and
time, allow users to choose between different social roles (e.g., friend, therapist etc.) so the PACA
can dynamically take on the user’s preferred social role, but can also switch between social roles
that promote the user’s therapy progress.
DP5 Principle of Anthropomorphism For designers and developers to design personality-adaptive conversational agents (PACA) that pro-
vide mental health services to socially support users independently of therapist, location and time,
allow the users to choose the type of PACA they want to interact with (chatbot, voice assistant,
embodied conversational agent), so they can determine the PACA’s degree of anthropomorphism
based on individual needs.
DP6 Principle of Personality Adaptivity For designers and developers to design personality-adaptive conversational agents (PACA) that
provide mental health services to socially support users independently of therapist, location and
time, imbue the PACA with language cues specific to different personality dimensions to enable
the PACA to adapt to the user’s preferred personality and increase interaction quality.
936
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Information Systems Frontiers (2022) 24:923–943
1 3
device and for purposes of testing” (Gregor etal., 2007, p.
322). For this purpose, we modeled a PACA starting from
the proposed DPs, so that system engineers and software
developers have the abstract design knowledge of the DPs
available in a transferable form. We opted for a graphical
modeling for the specification, design, documentation and
visualization of software components and interfaces to trans-
fer our DPs in applicable form. In our expository instantia-
tion, we decided to model the functionality of a PACA in a
technology-independent representation. Since technologies
change very quickly, especially in the field of AI (e.g., chat-
bot services), the model can serve as a blueprint for future
solutions. Figure3 shows the model diagram of a PACA for
mental health support.
Interaction with the PACA takes place via a conversa-
tional user interface (CUI), which ensures continuous avail-
ability for the user and the possibility to proactively con-
tact the user through push notifications, for example (DP1).
Examples of typical CUIs include messaging applications,
such as WhatsApp or Facebook Messenger, or a standalone
service implemented on a website or in a mobile application.
The CUI is also responsible for the representation of the
PACA and its anthropomorphic features (DP5). To be able to
display all cues desired by the user, the CUI should be able
to provide both text and speech-based output and should be
able to represent an embodied PACA (i.e., body language).
Therefore, integration in a website or as a mobile application
that allows a rich representation of a PACA is preferable.
The bot service is responsible for the logical interaction with
the user, and therefore is able to process natural language.
In addition, the bot service can adapt its conversational
design, which determines the style of communication and
personality adaptivity (DP6). The conversational design is
usually created by a conversational designer, who should
co-develop it with a therapist in the case of a PACA. The
conversational design is guided by the application logic,
which in turn draws on the database and the results of the
personality mining service. This functionality could be real-
ized with services such as Google Dialogflow, IBM Watson
Assistant, Amazon Lex, or Wit.AI.
The application is the core of the PACA and contains the
general logic of the system. It accesses information from
the knowledge base (database) and the information from the
personality insight service to decide how to interact with the
user. The application specifies not only the social role (DP4),
but also the type of communication (DP6) and the degree of
the anthropomorphism (DP5), which is the core of its adapt-
ability. Together with the knowledge base and the identified
personality of the user, it represents the competence of the
PACA (DP2), which is reflected in the conversational design
of the bot service and the CUI. The application should be
implemented as platform-independently as possible and be
able to communicate with the database, with the personal-
ity mining service, and with the bot service, which is why
technologies or programming languages such as JavaScript,
Python, Scala, Ruby, or Go are suitable. The personality
mining service is responsible for analyzing the user’s mes-
sages and determining the personality traits of the user. It
accesses the data that the user enters via the CUI, and ana-
lyzes it. The results are then passed on to the application.
One approach is to analyze specific word usage and build
statistical models that determine the personality traits of
the user. The foundation of these procedures was laid by
Pennebaker and Francis (1996) with the Linguistic Inquiry
Fig. 3 Expository instantiation
937
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Information Systems Frontiers (2022) 24:923–943
1 3
and Word Count (LIWC), which quickly gained promi-
nence in linguistic analysis (Chung & Pennebaker, 2012;
Pennebaker etal., 2015). Further improvements have been
achieved using so-called word embeddings, such as Google’s
word2vec or Stanford’s GloVe (Pennington etal., 2014). The
advantage of such models is that semantic similarity between
words is determined unsupervised, whereas LIWC relies on
human judges and psychologists to determine the meaning
of words (Arnoux etal., 2017; Rice & Zorn, 2021). Arnoux
etal. (2017) suggest that models with word embeddings pre-
dict personality even better than models based on LIWC.
IBM developed Watson Personality Insights, a commercial
software package, as a service for ready-to-use personality
predictions based on GloVe and Twitter posts (Arnoux etal.,
2017). These models or services can be used to determine
the personality traits of the user from their interactions. With
the user’s personality traits identified, the PACA can adapt to
the user according to the program logic (application).
The database contains the entire knowledge base of the
PACA, ranging from communication forms, anthropomor-
phic expressions, and other cues to therapeutic methods and
skills. The system also constantly learns from past interac-
tions with all its users, as the application not only accesses
the database, but also regularly updates it with informa-
tion. The database also represents the learning aspect of
a PACA, as it stores not only all therapeutic methods and
knowledge bases, but also the knowledge gained from past
interactions. For secure handling of the user’s sensitive data,
it is important that all data is protected with the latest tech-
nology (DP3), such as multiple factor authentication and
data encryption. This should not only be ensured across all
components, but also be communicated to the user via the
bot services, so that they can build trust and rapport with
the PACA (DP3). We previously published a paper that
describes the implementation of a PACA within another
application domain (Ahmad etal., 2020a).
6 Discussion
6.1 Theoretical andPractical Contributions
We contribute to the body of design knowledge for CAs
by providing evaluated DPs for the design of a PACA. The
six derived DPs can be divided into two categories. The
first category contains DP1: Principle of Proactive Sup-
port, DP2: Principle of Competence and DP3: Principle of
Transparency. These three principles represent the founda-
tion of a PACA and can be considered as basic prerequisites
when designing a PACA. Proactive support, competence
and transparency are particularly important design elements
in the context of mental health care, as emphasized in our
expert panel’s reports. However, it is the second category
of DPs that transforms a CA into a PACA: DP4: Principle
of Social Role, DP5: Principle of Anthropomorphism and
DP6: Principle of Personality Adaptivity are the princi-
ples that enable adaptation in the form of customization
and personalization to user preferences, and consequently
provide a more tailored service based on users’ individual
needs and personalities. As evaluated by the experts, DP6
in particular – which represents a necessary requirement
for a PACA – is a crucial design element. The DPs do not
build on each other and can be considered as stand-alone
design elements. However, we believe that only a combina-
tion of all DPs provides the best possible results, since the
six DPs cover all important criteria to design a PACA for
mental health care.
Our work also offers several practical contributions.
We generated design knowledge in the form of prescrip-
tive knowledge, and we provide guidance for CA design-
ers and developers. Based on our experts’ evaluation, we
argue that instantiating our DPs to design a PACA should
improve interaction with users who seek support for their
mental health care. While it is unlikely that PACAs that
mimic human interactions could ever completely replace
human psychotherapists, they may be a promising source of
support for a wide range of user groups and different situa-
tions. First, for users who want to receive social support as
an everyday social interaction to reduce loneliness or prevent
early mental health issues, PACAs can surpass availability
of and ease the burden on health care providers by being
accessible to provide counseling at any time. Especially in
the current COVID-19 pandemic, the need for social support
is of immense importance, and the lack of such support can
otherwise profoundly affect a person’s quality of life. Sec-
ond, for users who are undergoing therapy for more severe
mental health issues and are receiving treatment from human
professionals, a PACA may also be beneficial by deliver-
ing added value to therapeutic interventions. Since PACAs
are not susceptible to forgetfulness or fatigue, they can be
used as additional support systems for both the patient and
the provider, by offering periodic check-ups, for example.
In these cases, however, it is crucial for the PACA to be
monitored by human professionals to ensure patient safety,
as proposed by our experts.
938
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Information Systems Frontiers (2022) 24:923–943
1 3
6.2 Limitations andEthical Considerations
A number of limitations have to be considered with respect
to our qualitative study and expert interviews. A number of
our study participants indicated that they have no experi-
ence with mental health related therapies. Hence, the USs
we gathered would probably have led to different require-
ments if all participants had a similar amount of experi-
ence with mental health related therapies. However, as our
intention was to include a non-clinical population, that is
people who might want to receive social support as an eve-
ryday social interaction, we did not exclude participants
from our study who indicated that they have no experience
with mental health therapies. Another limitation is the small
number of experts who evaluated the DPs. All our experts
have less than 10years of experience in their field and are
about the same age. Experts with significantly more years
of experience (e.g., < 20years) might have evaluated the
DPs differently.
Despite the cost, reach, and scalability advantages of
CAs over human counsellors and therapists, it is important
to note that there may be several drawbacks to using CAs
in mental health care. First, these CAs could cause users
to isolate themselves even more from the outside world.
Since it can be easier to establish a relationship with these
CAs than with another human (e.g., a therapist), users may
lose interest in meeting or spending time with with their
human friends and family members (Skjuve etal., 2021).
This is even more problematic as many CAs are deliberately
designed to look and act like humans (e.g., by giving them
names and avatars), which further blurs the line between
humans and machines (Porra etal., 2019). Second, these
CAs are developed and operated by companies with busi-
ness interests. Therefore, some users worry that the sensitive
information they present in their conversations with the CAs
could be intentionally or unintentionally shared with third
parties (Bae Brandtzæg etal., 2021). Finally, as users have
no control over their own CA, they become very dependent
on the company that operates the CA, and some are afraid
that their CAs could be deleted at some point (Skjuve etal.,
2021). When taking all these considerations together, it is
important to keep these drawbacks in mind when designing
PACAs. Nevertheless, we believe that the potential advan-
tages outweigh the potential drawbacks when these CAs are
designed appropriately.
7 Conclusion
Though being a fruitful area with large practical potential,
the adoption of CAs for mental health care is associated
with some challenges. Literature issues involve the inability
of current CAs to capture and adapt dynamically to user
personality, as well as to provide responses and reactions
tailored in accordance with users’ individual characteris-
tics to an adequate extent. In addition, CAs do not live up
to their full potential yet, as they have been shown to have
limited conversational abilities, particularly when it comes
to longer and more complex interactions. These issues, how-
ever, are important factors that need to be taken into consid-
eration when addressing primary issues from the application
domain, such as the increased necessity for health-related
services, the shortage of health care providers, or the para-
mount importance of patient safety. Motivated by the lack
of design knowledge for PACAs, we proposed and evaluated
DPs for designing PACAs that improve interaction, specifi-
cally for users who seek support for their mental health. We
focused on two steps to answer our research question: Based
on our kernel theories we first identified current issues from
the AD and LIs, as well as derived USs, through a qualita-
tive study. Within this first step of our problem space, we
then derived MRs. On the basis of step one, and as part of
our solution space, we proposed DPs for PACAs in step two.
We then conducted expert interviews with psychologists and
psychiatrists to evaluate the derived DPs, and adjusted and
refined our final DPs based on their feedback. In a last step,
we transferred our defined DPs to an expository instantiation
for the purpose of better visualization.
According to the DSR contribution framework pro-
posed by Gregor and Hevner (2013), our work can be
classified as an improvement, as we address a known
problem with a new solution. We provide prescriptive
design knowledge by deriving and evaluating DPs for
PACAs in mental health care. Our DPs contribute to the
body of design knowledge for CAs and provide guidance
for practitioners, such as designers, developers, and men-
tal health organizations, on how to design PACAs that can
better support their users. Instantiating these principles
may improve interaction with users who seek support for
mental health issues. We believe that our design approach
could also be a valuable starting point for the design of
PACAs in other domains.
939
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Information Systems Frontiers (2022) 24:923–943
1 3
Appendix
Figure4
Funding Open Access funding enabled and organized by Projekt
DEAL.
Declarations
Conflict of Interest The authors have no conflicts of interest to declare.
All co-authors have seen and agree with the contents of the manu-
script, and there is no financial interest to report.
Open Access This article is licensed under a Creative Commons Attri-
bution 4.0 International License, which permits use, sharing, adapta-
tion, distribution and reproduction in any medium or format, as long
as you give appropriate credit to the original author(s) and the source,
provide a link to the Creative Commons licence, and indicate if changes
were made. The images or other third party material in this article are
included in the article's Creative Commons licence, unless indicated
otherwise in a credit line to the material. If material is not included in
the article's Creative Commons licence and your intended use is not
permitted by statutory regulation or exceeds the permitted use, you will
need to obtain permission directly from the copyright holder. To view a
copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.
References
Abd-alrazaq, A. A., Alajlani, M., Alalwan, A. A., Bewick, B. M.,
Gardner, P., & Househ, M. (2019). An overview of the features
of Chatbots in mental Health: A scoping review. International
Journal of Medical Informatics, 132, 103978 s.
Abd-Alrazaq, A. A., Alajlani, M., Ali, N., Denecke, K., Bewick, B.
M., & Househ, M. (2021). Perceptions and opinions of patients
about mental Health Chatbots: Scoping review. Journal of Medi-
cal Internet Research, 23(1), e17828.
Fig. 4 Mapping diagram of the derivation and construction of the MRs and DPs (Möller etal., 2020)
940
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Information Systems Frontiers (2022) 24:923–943
1 3
Ahmad, R., Siemon, D., Fernau, D., & Robra-Bissantz, S. (2020a).
“Introducing" Raffi": A Personality Adaptive Conversational
Agent.,” in PACIS, p. 28.
Ahmad, R., Siemon, D., & Robra-Bissantz, S. (2020b). ExtraBot vs
IntroBot: The Influence of Linguistic Cues on Communication
Satisfaction. In B. B. Anderson, J. Thatcher, R. D. Meservy, K.
Chudoba, K. J. Fadel, & S. Brown (Eds.), 26th Americas Confer-
ence on Information Systems, AMCIS 2020, Virtual Conference,
August 15–17, 2020. Association for Information Systems.
Ahmad, R., Siemon, D., Gnewuch, U., & Robra-Bissantz, S. (2021a).
The benefits and caveats of personality-adaptive conversational
agents in mental Health care. AMCIS.
Ahmad, R., Siemon, D., & Robra-Bissantz, S. (2021b). “Communi-
cating with machines: Conversational agents with personality
and the role of extraversion,” in Proceedings of the 54th Hawaii
International Conference on System Sciences, p. 4043.
Allport, G. W. (1961). Pattern and growth in personality.
Al-Natour, S., Benbasat, I., & Cenfetelli, R. T. (2005). “The Role of
Similarity in E-Commerce Interactions: The Case of Online
Shopping Assistants,” SIGHCI 2005 Proceedings, p. 4.
Arnoux, P.-H., Xu, A., Boyette, N., Mahmud, J., Akkiraju, R., & Sinha,
V. (2017). “25 Tweets to Know You: A New Model to Predict
Personality with Social Media,” in Proceedings of the Interna-
tional AAAI Conference on Web and Social Media (Vol. 11).
Babbie, E. R. (2020). The practice of social research, Cengage learning.
Bae Brandtzæg, P. B., Skjuve, M., Kristoffer Dysthe, K. K., & Føl-
stad, A. (2021). “When the social becomes non-human: Young
People’s perception of social support in Chatbots,” in Proceed-
ings of the 2021 CHI Conference on Human Factors in Com-
puting Systems, pp. 1–13.
Baskerville, R., & Pries-Heje, J. (2010). Explanatory design theory.
Business & Information Systems Engineering, 2(5), 271–282.
Bendig, E., Erb, B., Schulze-Thuesing, L., & Baumeister, H. (2019).
The next generation: Chatbots in clinical psychology and psy-
chotherapy to Foster mental Health – A scoping review (pp.
1–13). Karger Publishers.
Bouchet, F., & Sansonnet, J.-P. (2012). “Intelligent Agents with Per-
sonality: From Adjectives to Behavioral Schemes,” in Cog-
nitively Informed Intelligent Interfaces: Systems Design and
Development, IGI Global, pp. 177–200.
Boyd, R. L., & Pennebaker, J. W. (2017). Language-based personal-
ity: A new approach to personality in a digital world. Current
Opinion in Behavioral Sciences, 18, 63–68.
Brendel, A. B., Mirbabaie, M., Lembcke, T.-B., & Hofeditz, L.
(2021). Ethical management of artificial intelligence. Sustain-
ability, 13(4), 1974.
Chakrabarti, C., & Luger, G. F. (2015). Artificial conversations for cus-
tomer service chatter bots: Architecture, algorithms, and evaluation
metrics. Expert Systems with Applications, 42(20), 6878–6897.
Chung, C. K., & Pennebaker, J. W. (2012). Linguistic Inquiry and Word
Count (LIWC): Pronounced ‘Luke,’... and Other Useful Facts.
InApplied Natural Language Processing: Identification, Investi-
gation and Resolution (pp. 206–229).
Clark, L., Pantidi, N., Cooney, O., Doyle, P., Garaialde, D., Edwards,
J., Spillane, B., Gilmartin, E., Murad, C., Munteanu, C., Wade,
V., & Cowan, B. R. (2019). “What Makes a Good Conversation?:
Challenges in Designing Truly Conversational Agents,” in Pro-
ceedings of the 2019 CHI Conference on Human Factors in Com-
puting Systems, Glasgow Scotland Uk: ACM, May 2, pp. 1–12.
D’Alfonso, S. (2020). AI in mental Health. Current Opinion in Psychol-
ogy, 36, 112–117.
Diederich, S., Brendel, A., Morana, S., & Kolbe, L. (2022). On the
design of and interaction with conversational agents: an organ-
izing and assessing review of human-computer interaction
research. Journal of the Association for Information Systems,
23(1), 96–138.
Feine, J., Gnewuch, U., Morana, S., & Maedche, A. (2019). A tax-
onomy of social cues for conversational agents. International
Journal of Human-Computer Studies, 132, 138–161.
Ferrucci, D. A. (2012). Introduction to ‘This Is Watson. IBM Journal
of Research and Development, 56(3.4), 1–1.
Fiske, A., Henningsen, P., & Buyx, A. (2019). Your robot therapist will
see you now: Ethical implications of embodied artificial intel-
ligence in psychiatry, psychology, and psychotherapy. Journal of
Medical Internet Research, 21(5)). https:// doi. org/ 10. 2196/ 13216
Fogg, B. J. (2002). “Persuasive Technology: Using Computers to
Change What We Think and Do,” Ubiquity (2002:December),
ACM New York, NY, USA, p. 2.
Gaffney, H., Mansell, W., & Tai, S. (2019). Conversational Agents in
the Treatment of Mental Health Problems: Mixed-Method Sys-
tematic Review. JMIR Mental Health, 6(10), e14166.
Gnewuch, U., Morana, S., & Maedche, A. (2017). "Towards design-
ing cooperative and social conversational agents for customer
service," in Proceedings of the 38th International Conference on
Information Systems (ICIS2017).
Gnewuch, U., Yu, M., & Maedche, A. (2020). “The Effect of Perceived
Similarity in Dominance on Customer Self-Disclosure to Chat-
bots in Conversational Commerce,” in Proceedings of the 28th
European Conference on Information Systems (ECIS 2020).
Golbeck, J., Robles, C., Edmondson, M., & Turner, K. (2011). “Pre-
dicting Personality from Twitter,” in Privacy, Security, Risk and
Trust (PASSAT) and 2011 IEEE Third Inernational Conference
on Social Computing (SocialCom), 2011 IEEE Third Interna-
tional Conference On, IEEE, pp. 149–156.
Goldberg, L. R. (1993). “The Structure of Phenotypic Personality
Traits.,” American Psychologist (48:1), American Psychologi-
cal Association, p. 26.
Graham, S., Depp, C., Lee, E. E., Nebeker, C., Tu, X., Kim, H.-C.,
& Jeste, D. V. (2019). Artificial intelligence for mental Health
and mental illnesses: An overview. Current Psychiatry Reports,
21(11), 116.
Graham, S. A., Lee, E. E., Jeste, D. V., Van Patten, R., Twamley, E.
W., Nebeker, C., Yamada, Y., Kim, H.-C., & Depp, C. A. (2020).
Artificial intelligence approaches to predicting and detecting cog-
nitive decline in older adults: A conceptual review. Psychiatry
Research, 284, 112732.
Gregor, S., & Hevner, A. R. (2013). Positioning and presenting design
science research for maximum impact. MIS Q, 37(2), 337–356.
Gregor, S., Jones, D., & etal. (2007). The Anatomy of a Design The-
ory, Association for Information Systems.
Gregor, S., Chandra Kruse, L., & Seidel, S. (2020). Research perspec-
tives: The anatomy of a design principle. Journal of the Associa-
tion for Information Systems, 21(6), 2.
Grudin, J., & Jacques, R. (2019). “Chatbots, Humbots, and the quest for
artificial general intelligence,” in Proceedings of the 2019 CHI
Conference on Human Factors in Computing Systems - CHI ‘19,
Glasgow, Scotland Uk: ACM Press, pp. 1–11.
Grünzig, S.-D., Baumeister, H., Bengel, J., Ebert, D., & Krämer, L.
(2018). “Effectiveness and Acceptance of a Web-Based Depres-
sion Intervention during Waiting Time for Outpatient Psycho-
therapy: Study Protocol for a Randomized Controlled Trial,”
Trials (19:1), Springer, pp. 1–11.
Hevner, A. R. (2007). A three cycle view of design science research.
Scandinavian Journal of Information Systems, 19(2), 4.
Hevner, A. R. (2020). The duality of science: Knowledge in infor-
mation systems research. Journal of Information Technology,
0268396220945714.
Hevner, A., March, S. T., Park, J., & Ram, S. (2004). Design science
research in information systems. MIS Quarterly, 28(1), 75–105.
Iivari, J. (2015). Distinguishing and contrasting two strategies for
design science research. European Journal of Information Sys-
tems, 24(1), 107–115.
941
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Information Systems Frontiers (2022) 24:923–943
1 3
Iivari, J., Hansen, M. R. P., & Haj-Bolouri, A. (2021). A proposal for
minimum reusability evaluation of design principles. European
Journal of Information Systems, 30(3), 286–303.
Jones, S. P., Patel, V., Saxena, S., Radcliffe, N., Ali Al-Marri, S.,
& Darzi, A. (2014). How google’s ‘ten things we know to be
true’could guide the development of mental health mobile apps.
Health Affairs, 33(9), 1603–1611.
Junglas, I. A., Johnson, N. A., & Spitzmüller, C. (2008). Personality
traits and concern for privacy: An empirical study in the context
of location-based services. European Journal of Information
Systems, 17(4), 387–402.
Kampman, O., Siddique, F. B., Yang, Y., & Fung, P. (2019). Adapting a
Virtual Agent to User Personality. InAdvanced Social Interaction
with Agents (pp. 111–118). Springer.
Kerr, I. R. (2003). Bots, babes and the Californication of commerce.
University of Ottawa Law and Technology Journal, 1, 285.
Kim, S. Y., Schmitt, B. H., & Thalmann, N. M. (2019). Eliza in the Uncanny
Valley: Anthropomorphizing consumer robots increases their per-
ceived warmth but decreases liking. Marketing Letters, 30(1), 1–12.
Kocaballi, A. B., Berkovsky, S., Quiroz, J. C., Laranjo, L., Tong, H.
L., Rezazadegan, D., Briatore, A., & Coiera, E. (2019). The per-
sonalization of conversational agents in Health care: Systematic
review. Journal of Medical Internet Research, 21(11), e15360.
Kocaballi, A. B., Laranjo, L., Quiroz, J., Rezazadegan, D., Kocielnik,
R., Clark, L., Liao, V., Park, S., Moore, R., & Miner, A. (2020).
Conversational agents for Health and wellbeing.
Laranjo, L., Dunn, A. G., Tong, H. L., Kocaballi, A. B., Chen, J.,
Bashir, R., Surian, D., Gallego, B., Magrabi, F., Lau, A. Y. S.,
& Coiera, E. (2018). Conversational agents in healthcare: A sys-
tematic review. Journal of the American Medical Informatics
Association, 25(9), 1248–1258.
Liu, K., & Picard, R. W. (2005). “Embedded Empathy in Continuous,
Interactive Health Assessment,” in CHI Workshop on HCI Chal-
lenges in Health Assessment (Vol. 1), Citeseer, p. 3.
Luxton, D. D. (2014). Recommendations for the ethical use and Design
of Artificial Intelligent Care Providers. Artificial Intelligence in
Medicine, 62(1), 1–10.
Luxton, D. D. (2020). Ethical implications of conversational agents in
global public Health. Bulletin of the World Health Organization,
98(4), 285.
Maedche, A., Gregor, S., Morana, S., & Feine, J. (2019). Conceptual-
ization of the problem space in design science research. InInter-
national Conference on Design Science Research in Information
Systems and Technology (pp. 18–31). Springer.
Mairesse, F., & Walker, M. A. (2010). Towards personality-based user
adaptation: Psychologically informed stylistic language generation.
User Modeling and User-Adapted Interaction, 20(3), 227–278.
Mairesse, F., Walker, M. A., Mehl, M. R., & Moore, R. K. (2007).
Using linguistic cues for the automatic recognition of personal-
ity in conversation and text. Journal of Artificial Intelligence
Research, 30, 457–500.
Mayring, P. (2014). Qualitative content analysis: Theoretical Founda-
tion, Basic Procedures and Software Solution.
McCrae, R. R., & Costa, P. T., Jr. (1997). Personality trait structure as
a human universal. American Psychologist, 52(5), 509.
McCrae, R. R., & John, O. P. (1992). An introduction to the five-factor
model and its applications. Journal of Personality, 60(2), 175–215.
McTear, M., Callejas, Z., & Griol, D. (2016). The conversational Inter-
face: Talking to smart devices. Springer.
Möller, F., Guggenberger, T. M., & Otto, B. (2020). Towards a method
for design principle development in information systems. InInter-
national Conference on Design Science Research in Information
Systems and Technology (pp. 208–220). Springer.
Moon, Y., & Nass, C. (1996). How ‘real’ are computer personalities?
Psychological responses to personality types in human-computer
interaction. Communication Research, 23(6), 651–674.
Nass, C., & Moon, Y. (2000). Machines and mindlessness: Social
responses to computers. Journal of Social Issues, 56(1), 81–103.
Nass, C., Steuer, J., Tauber, E., & Reeder, H. (1993). “Anthropomor-
phism, Agency, and Ethopoeia: Computers as Social Actors,” in
INTERACT ‘93 and CHI ‘93 Conference Companion on Human
Factors in Computing Systems, CHI ‘93, New York, NY, USA:
Association for Computing Machinery, April 1, pp. 111–112.
https:// doi. org/ 10. 1145/ 259964. 260137.
Nass, C., Steuer, J., & Tauber, E. R. (1994). “Computers are social
actors,” in Proceedings of the SIGCHI Conference on Human
Factors in Computing Systems, ACM, pp. 72–78.
Nass, C., Moon, Y., Fogg, B. J., Reeves, B., & Dryer, D. C. (1995). Can
computer personalities be human personalities? International
Journal of Human-Computer Studies, 43(2), 223–239.
Natale, S. (2018). If software is narrative: Joseph Weizenbaum, arti-
ficial intelligence and the biographies of ELIZA. New Media &
Society, 21, 146144481880498.
Nißen, M. K., Selimi, D., Janssen, A., Cardona, D. R., Breitner, M. H.,
Kowatsch, T., & von Wangenheim, F. (2021). See you soon again,
Chatbot? A design taxonomy to characterize user-Chatbot relationships
with different time horizons. Computers in Human Behavior, 107043.
Pennebaker, J. W. (2011). The secret life of pronouns: How our words
reflect who we are. Bloomsbury.
Pennebaker, J. W., & Francis, M. E. (1996). Cognitive, emotional, and
language processes in disclosure. Cognition & Emotion, 10(6)
Taylor & Francis, 601–626.
Pennebaker, J. W., Boyd, R. L., Jordan, K., & Blackburn, K. (2015).
The development and psychometric properties of LIWC2015.
Pennington, J., Socher, R., & Manning, C. (2014). “Glove: Global vec-
tors for word representation,” in Proceedings of the 2014 Con-
ference on Empirical Methods in Natural Language Processing
(EMNLP), pp. 1532–1543.
Peters, O. (2013). Critics of Digitalisation: Against the Tide: Warners,
Sceptics, Scaremongers, Apocalypticists: 20 Portraits, Studien
Und Berichte Der Arbeitsstelle Fernstudienforschung Der Carl
von Ossietzky Universität Oldenburg, Oldenburg: BIS-Verlag der
Carl von Ossietzky Universität Oldenburg.
Porra, J., Lacity, M., & Parks, M. S. (2019). “‘Can computer based
human-likeness endanger humanness?’ – A philosophical and
ethical perspective on digital assistants expressing feelings they
Can’t have””, Information Systems Frontiers.
Prakash, A. V., & Das, S. (2020). Intelligent conversational agents in mental
healthcare services: A thematic analysis of user perceptions. Pacific
Asia Journal of the Association for Information Systems, 12(2), 1.
Purao, S., Chandra Kruse, L., & Maedche, A. (2020). The origins of
design principles: Where do… they all come from?
Ranjbartabar, H., Richards, D., Kutay, C., & Mascarenhas, S. (2018).
“Sarah the virtual advisor to reduce study stress,” In Proceedings
of the 17th International Conference on Autonomous Agents and
MultiAgent Systems, pp. 1829–1831.
Rice, D. R., & Zorn, C. (2021). Corpus-based dictionaries for sen-
timent analysis of specialized vocabularies. Political Science
Research and Methods, 9(1) Cambridge University Press, 20–35.
Rothe, H., Wessel, L., & Barquet, A. P. (2020). Accumulating design
knowledge: A mechanisms-based approach. Journal of the Asso-
ciation for Information Systems, 21(3), 1.
Schuetzler, R., Grimes, G., Giboney, J., & Nunamaker, J. (2018). “The
influence of conversational agents on socially desirable respond-
ing,” Proceedings of the 51st Hawaii International Conference on
System Sciences, pp. 283–292.
Shah, H., Warwick, K., Vallverdú, J., & Wu, D. (2016). Can machines
talk? Comparison of Eliza with modern dialogue systems. Com-
puters in Human Behavior, 58, 278–295.
Shum, H., He, X., & Li, D. (2018). From Eliza to XiaoIce: Challenges
and opportunities with social Chatbots. Frontiers of Information
Technology & Electronic Engineering, 19(1), 10–26.
942
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Information Systems Frontiers (2022) 24:923–943
1 3
Skjuve, M., Følstad, A., Fostervold, K. I., & Brandtzaeg, P. B. (2021).
My Chatbot companion-a study of human-Chatbot relationships.
International Journal of Human-Computer Studies, 149), Else-
vier, 102601.
Smith, K., Masthoff, J., Tintarev, N., & Wendy, M. (2015). Adapting
emotional support to personality for Carers experiencing. Stress.
https:// doi. org/ 10. 13140/ RG.2. 1. 3898. 9929
Stieger, M., Nißen, M., Rüegger, D., Kowatsch, T., Flückiger, C., and
Allemand, M. 2018. “PEACH, a smartphone-and conversational
agent-based coaching intervention for intentional personality
change: Study protocol of a randomized, wait-list controlled
trial,” BMC Psychology (6:1), Springer, pp. 1–15.
Ta, V., Griffith, C., Boatfield, C., Wang, X., Civitello, M., Bader, H.,
DeCero, E., & Loggarakis, A. (2020). User experiences of social
support from companion Chatbots in everyday contexts: The-
matic analysis. Journal of Medical Internet Research, 22(3),
e16235.
Torous, J., Myrick, K. J., Rauseo-Ricupero, N., & Firth, J. (2020).
Digital mental health and COVID-19: Using technology today
to accelerate the curve on access and quality tomorrow. JMIR
Mental Health, 7(3), e18848. https:// doi. org/ 10. 2196/ 18848
Venable, J. (2006). “The Role of Theory and Theorising in Design Sci-
ence Research,” in Proceedings of the 1st International Confer-
ence on Design Science in Information Systems and Technology
(DESRIST 2006), Citeseer, pp. 1–18.
Völkel, S. T., Kempf, P., & Hussmann, H. (2020). “Personalised chats
with voice assistants: The user perspective,” in Proceedings of
the 2nd Conference on Conversational User Interfaces, pp. 1–4.
Völkel, S. T., Meindl, S., & Hussmann, H. (2021). “Manipulating and
evaluating levels of personality perceptions of voice assistants
through enactment-based dialogue design,” in CUI 2021-3rd
Conference on Conversational User Interfaces, pp. 1–12.
Vom Brocke, J., Winter, R., Hevner, A., & Maedche, A. (2020). Special
issue editorial–accumulation and evolution of design knowledge
in design science research: A journey through time and space.
Journal of the Association for Information Systems, 21(3), 9.
Wasil, A. R., Palermo, E., Lorenzo-Luaces, L., & DeRubeis, R. (2021).
Is there an app for that? A review of popular mental health and
wellness apps, PsyArXiv. https:// doi. org/ 10. 31234/ osf. io/ su4ar.
Weizenbaum, J. (1966). ELIZA—A Computer program for the study
of natural language communication between man and machine.
Communications of the ACM, 9(1), 36–45.
WHO. (2017). Depression and other common mental disorders: Global
Health estimates. World Health Organization.
WHO. (2021). “WHO Executive Board Stresses Need for Improved
Response to Mental Health Impact of Public Health Emergen-
cies”. https:// www. who. int/ news/ item/ 11- 02- 2021- who- execu
tive- board- stres ses- need- for- impro ved- respo nse- to- mental-
health- impact- of- public- health- emerg encies. Accessed April
20, 2021.
Wibhowo, C., & Sanjaya, R. (2021). “Virtual assistant to suicide pre-
vention in individuals with borderline personality disorder,” in
2021 International Conference on Computer & Information Sci-
ences (ICCOINS), IEEE, pp. 234–237.
Woebot Health. (2021). “Woebot”. https:// woebo theal th. com/ produ
cts- pipel ine/. Accessed Feb 28, 2021.
Wysa, io. (2021). “Wysa”. https:// www. wysa. io/.Accessed Feb 28,
2021.
X2, A. (2021). “Tess”. https:// www. x2ai. com/.Accessed Feb 28, 2021.
Yarkoni, T. (2010). Personality in 100,000 words: A large-scale anal-
ysis of personality and word use among bloggers. Journal of
Research in Personality, 44(3), 363–373.
Yorita, A., Egerton, S., Oakman, J., Chan, C., & Kubota, N. (2019).
“Self-adapting Chatbot personalities for better peer support,” in
2019 IEEE international conference on systems, Man and Cyber-
netics (SMC), IEEE, pp. 4094–4100.
Zalake, M. (2020). “Advisor: Agent-based intervention leveraging
individual differences to support mental wellbeing of college
students,” in Extended Abstracts of the 2020 CHI Conference on
Human Factors in Computing Systems, pp. 1–8.
Publisher’s Note Springer Nature remains neutral with regard to
jurisdictional claims in published maps and institutional affiliations.
Rangina Ahmad received her B.Sc. and M.Sc. degrees in business
information systems from the Technische Universität Braunschweig
in Germany, where she is currently pursuing her Ph.D. degree. Since
2018, she has been a Research Associate at the Chair of Information
Management at the Technische Universität Braunschweig. Her research
focuses on topics such as human-AI interaction, personality psychol-
ogy, and e-services. Her work has been published in leading informa-
tion systems conferences, such as Americas Conference on Information
Systems and Hawaii International Conference on Information Systems
Dominik Siemon is an Associate Professor with the Department of
Software Engineering, Lappeenranta-Lahti University of Technology
(LUT University), Finland. He studied business information systems
at the Technische Universität Braunschweig in Germany, where he
received his Dr. rer. pol. (Ph.D.) degree in business information sys-
tems. He worked as a Research Assistant and a Postdoctoral Researcher
at the Technische Universität Braunschweig and a part-time Professor
of business information systems and digital business at IU International
University. His mainly design-oriented research in the field of informa-
tion systems addresses collaboration and interaction with intelligent
systems, conversational agents, innovation management, collaboration
technology, and creativity. His work has been published in leading con-
ferences, such as the International Conference on Information Systems,
and in journals, such as Education and Information Technologies, AIS
Transactions on Human-Computer Interaction, and the Communica-
tions of the Association for Information Systems
Ulrich Gnewuch is a Postdoctoral Researcher at the Institute of Infor-
mation Systems and Marketing at Karlsruhe Institute of Technology
(KIT), Germany. He studied Information Systems at the University of
Mannheim and received his PhD degree from KIT. His research focuses
on the design of conversational user interfaces and digital assistants.
His research has been published in journals such as the International
Journal of Human-Computer Studies and Computers in Human Behav-
ior. He serves as the vice chair for teaching resources of the AIS Spe-
cial Interest Group on Human-Computer Interaction (SIGHCI)
Susanne Robra-Bissantz is Professor for Information Management and
has been head of the Institute of Information Systems and the Chair of
Information Management at the Technische Universität Braunschweig
in Germany since 2007. After receiving her doctorate in Economic and
Social Sciences, she worked as a Research Assistant and habilitated
at the Chair of Business Administration at the Friedrich–Alexander
University Erlangen–Nürnberg, Germany. As Vice President for Stud-
ies and Cooperation, she actively worked on new forms of teaching
and examination and has implemented numerous third-party funded
projects in cooperation with companies. Her design-oriented research
focuses on holistic digital service development, e-collaboration and
smart participation. She published her research at international confer-
ences and in recognized journals
943
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

1.
2.
3.
4.
5.
6.
Terms and Conditions
Springer Nature journal content, brought to you courtesy of Springer Nature Customer Service Center GmbH (“Springer Nature”).
Springer Nature supports a reasonable amount of sharing of research papers by authors, subscribers and authorised users (“Users”), for small-
scale personal, non-commercial use provided that all copyright, trade and service marks and other proprietary notices are maintained. By
accessing, sharing, receiving or otherwise using the Springer Nature journal content you agree to these terms of use (“Terms”). For these
purposes, Springer Nature considers academic use (by researchers and students) to be non-commercial.
These Terms are supplementary and will apply in addition to any applicable website terms and conditions, a relevant site licence or a personal
subscription. These Terms will prevail over any conflict or ambiguity with regards to the relevant terms, a site licence or a personal subscription
(to the extent of the conflict or ambiguity only). For Creative Commons-licensed articles, the terms of the Creative Commons license used will
apply.
We collect and use personal data to provide access to the Springer Nature journal content. We may also use these personal data internally within
ResearchGate and Springer Nature and as agreed share it, in an anonymised way, for purposes of tracking, analysis and reporting. We will not
otherwise disclose your personal data outside the ResearchGate or the Springer Nature group of companies unless we have your permission as
detailed in the Privacy Policy.
While Users may use the Springer Nature journal content for small scale, personal non-commercial use, it is important to note that Users may
not:
use such content for the purpose of providing other users with access on a regular or large scale basis or as a means to circumvent access
control;
use such content where to do so would be considered a criminal or statutory offence in any jurisdiction, or gives rise to civil liability, or is
otherwise unlawful;
falsely or misleadingly imply or suggest endorsement, approval , sponsorship, or association unless explicitly agreed to by Springer Nature in
writing;
use bots or other automated methods to access the content or redirect messages
override any security feature or exclusionary protocol; or
share the content in order to create substitute for Springer Nature products or services or a systematic database of Springer Nature journal
content.
In line with the restriction against commercial use, Springer Nature does not permit the creation of a product or service that creates revenue,
royalties, rent or income from our content or its inclusion as part of a paid for service or for other commercial gain. Springer Nature journal
content cannot be used for inter-library loans and librarians may not upload Springer Nature journal content on a large scale into their, or any
other, institutional repository.
These terms of use are reviewed regularly and may be amended at any time. Springer Nature is not obligated to publish any information or
content on this website and may remove it or features or functionality at our sole discretion, at any time with or without notice. Springer Nature
may revoke this licence to you at any time and remove access to any copies of the Springer Nature journal content which have been saved.
To the fullest extent permitted by law, Springer Nature makes no warranties, representations or guarantees to Users, either express or implied
with respect to the Springer nature journal content and all parties disclaim and waive any implied warranties or warranties imposed by law,
including merchantability or fitness for any particular purpose.
Please note that these rights do not automatically extend to content, data or other material published by Springer Nature that may be licensed
from third parties.
If you would like to use or distribute our Springer Nature journal content to a wider audience or on a regular basis or in any other manner not
expressly permitted by these Terms, please contact Springer Nature at
onlineservice@springernature.com